Feb., 1940
ELECTROSTATIC INFLUENCE OF
nitrate solution using a dichlorofluorescein adsorption indicator according to Kolthoff and collaborators.lO This method is applicable in the presence of silica, and has proved very efficient. We have not found the Volhard procedure t o be satisfactory in this work. Finally the several products obtained by the fluorination of the ethyl chloride were analyzed and results are summarized in Table 11. TABLE I1 ANALYTICAL VALUES Calcd. Compd.
F
Found
c1
F
CFzC12" 31.4 58.7 54.5 34.0 CFsCl 61.5 23.0 CFsCFzCl CFz=CClt 28.6 53.4 CHF,CH2Clb 37.8 35.3 a Analysis of purified Freon, * Decomposed in a Parr bomb.
31.4 . . . 54.6 54.3 61.4 61.5 28.6 28.6 37.8 38.0 to test the
c1
58.7 . . . 34.0 34.0 23.1 23.1 53.3 53.6 35.2 . . . procedure.
(10) Kolthoff, Lauer and Sunde, THISJOURNAL 61, 3273 (1929).
[CONTRIBUTION FROM
THE
SUBSTITUENTS ON
REACTION RATES
269
These results leave no doubt as to the composition of the compounds which were isolated in this work.
Summary Ethyl chloride was fluorinated in the vapor phase over copper gauze under varying conditions, and the relative amounts of the more important products determined. The compounds CF4, CF3C1, CFsCF2C1, CF2= CClz and CHFzCHzCl were isolated from the reaction mixture. The last three of these have not been obtained before by the action of elementary fluorine on an organic compound. A precise analytical procedure for determining fluorine and chlorine in stable organic gases has been described. DURHAM, NORTHCAROLINA RECBIVED AUGUST4, 1939
GEORGE HERBERT JONES LABORATORY OF THE UNIVERSITY O F CHICAGO]
The Electrostatic Influence of Substituents on Reactions Rates. I BY F. H. WESTHEIMER AND MARTINW. SHOOKHOFF Ingold has postulated that the effect of a polar substituent on the rate of saponification of an aliphatic ester, like the effect of such a substituent on the ionization constant of an acid, is largely electrostatic. In 19301 he employed Bjerrum's quantitative electrostatic formulation2 to compute, from the ratios of the rates of saponification of the first and second ester groups, the separation of the carbalkoxy groups in a series of esters of dibasic acids. Ingold's evidence went far toward establishing his hypothesis in those cases in which the molecule is long and the substituent can be regarded as a negative charge. Kirkwood and Westheimers have recently refined Bjerrum's approximate computations of the electrostatic free energy involved in a chemical reaction; the success of the new equations, while general, has been most conspicuous in the considerations of the ionization constants of short dibasic and of dipole substituted acidsJ4 cases for which the older formulation is inadequate. (1) Ingold, J. Chcm. Soc., 1375 (19301, 2170 (1931); Ingold and Mohrhenn, ibid., 1482 (1935). (2) Bjerrum, Z.fihysik. Chem., 106, 219 (1923). (31 Kirkwood and Westheimer, J. Chem. Phys., 6, 5013 (1938); Westheimer and Kirkwood, ibid., 6, 513 (1938). (4) Westheimer and Shookhoff, THIS JOURNAL, 61, 555 (1939); Westheimer, ibid., 61, 1977 (1939); cf. Eucken, Z . wrgew. Chcm., 4S, 201 (1932).
The shortcomings of Ingold's modified equation6 have been presented previously. The present paper deals with the rate of alkaline hydrolysis of esters and amides in cases in which the substituent is a negative charge, a positive charge and a dipole, all placed close to the seat of the reaction. When the substituent is close to the ester or amide bond that is broken, the rates of hydrolysis of the substituted and unsubstituted compounds differ by a factor of a hundred or more. Since the measurements were made b y conductivity, very slow and very fast reactions were excluded : the former because of action of alkali on the glass of the cell, the latter because of the time required to attain thermal equilibrium. These considerations necessitated a careful selection of the pairs of compounds employed. To illustrate the effect of a negative charge as substituent, the rate of alkaline hydrolysis of the sodium salt of oxamic acid was compared with the rate for oxamide; to illustrate the effect of a positive charge, the rate of saponification of the chloride of the tertiary butyl ester of betaine was compared with the rate for the tertiary butyl ester of dimethylglycine; and to illustrate the (5) Ingold, J. Chcm. Soc., 2179 (1931).
F. H. WESTHEIMER AND MARTINW. SHOOKHOFF
270
effect of a dipole, the rate of saponification of the tertiary butyl ester of chloroacetic acid was compared with the rate for the tertiary butyl ester of acetic acid. The equations of Kirkwood and Westheimer apply accurately only to spheres and to prolate ellipsoids in which substituents and charges are a t the foci. Due to the experimental limitations mentioned above, some sacrifice in the shape of the molecule was unavoidable. Theoretical Ingold treated the reaction velocity problem by means of Christiansen's theory,6 that is, as a case in which the local concentration of hydroxide ions differs, due to the presence in the molecule of a polar substituent, from that in the remainder of the solution. The assumptions involved can be more conveniently stated, however, when the problem is formulated in Bronsted's terms. Brijnsted postulated a rapid equilibrium between the reactants and an activated complex, followed by a slow decomposition of the activated complex to form the reaction products. In order to apply the electrostatic theory, it is further necessary to assume that the rates of decomposition will be the same for activated complexes formed from two closely similar molecules.* Stated mathematically kubJd.
= kK and
krobsd.
=
k'K'
and k'obsd, are the observed velocity constants, corrected to infinite dilution, for the bimolecular reactions, and K and K' are the equilibrium constants for the formation of the two activated complexes. If k and k', the rate constants for the decomposition of the activated complexes, are assumed equal, then kobsd,
kobsd./k'ubsd.
=
K/K'
In other words, to compute the ratio of the velocity constants a t infinite dilution, it is only necessary to compute the ratio of the equilibrium constants for the formation of the critical complexes. The rest of the argument exactly parallels that given by Kirkwood and Westheimer3 for the case of the ionization constants of acids with and without polar substituents. k"T log, K k"T log, K' is equal to Awl, where k" is Boltz(6) Christiansen, Z . physik. Chem., 113, 35 (1924). (7) Bronsted, Chem. Rev., 6 , 231 (1928). (8) The application of electrostatic theory t o the problem of the change of rate with solvent involves a more limited assumption (Scatchard, J . Chem. P h y s . , 1, 657 (1939)). Since the transmission coefficientis usually in the neighborhood of unity, a broader assumption is involved in Eyring's theory ( i b i d . , 3, 107 (1935)).
VOl. 62
mann's constant, T the absolute temperature, and Awl is the average reversible work expended in the transfer of a hydroxide ion from the unsubstituted activated complex to the substituted ester, infinitely separated in the solvent. Part of Awl is an entropy term, arising from differences in molecular symmetry. The electrostatic part of Aw', called Aw,is that which can be computed. The part of Awl connected with the intrinsic structure of the molecules is by no means to be overlooked, but i t is assumed that, as a first approximation, i t will vanish because of cancellations between the substituted and unsubstituted compounds. It follows, therefore, that AW =
k"T log, K/uK' = k"T log, k/uk' = 2.303 k"T A log k
where k" is the Boltzmann constant, is a statistical factor (equal to two in the saponification of esters or amides of dibasic acids) and the equation serves to define A log k. From the equations of Kirkwood and Westheimer and from the experimental values of A log k, i t is possible to compute the distance between the negative charge in the activated complexes, and the positive charge in the chloride of tertiary butyl betaine, the center of the dipole in tertiary butyl chloroacetate, and a positive charge in oxamide, located approximately in the position of the ionizable proton in oxamic acid. These distances are then compared with those obtained from an independent source. Experimental Apparatus.-Conductivity was measured with a calibrated Leeds and Northrup bridge, using a slide wire of the Kohlrausch type, and a Clough-Brengle audio frequency oscillator as the source of current. The apparatus was slightly modifiedQfrom the standard Washburn type, and was capable of measuring resistances with a precision of 0.01%. The design of the conductivity cells was dictated primarily by the experimental observation that the percentage time rate of change of conductivity due to the solution of the glass in alkali increases with temperature, dilution, and the ratio of surface to volume, as shown in Table I. Cell A, used at alkali concentrations of 0.0003 to 0.003 was spherical, of one liter capacity, and had bright electrodes, 1.8 cm. in diameter and 2 cm. apart. Due to the large size of this cell, a t least an hour was necessary to attain temperature equilibrium; it was not, then, very satisfactory for the more rapid reactions. For the concentration range 0.003 to 0.05, cell B was used. I t was cylindrical in shape, 3 cm. in diameter and 11 cm. (9) Jones and Josephs, THISJ O U R N A L , 60, 1049 (1828).
Feb., 1940
ELECTROSTATIC INFLUENCE OF SUBSTITUENTS ON REACTION RATES
27 1
long, with lightly platinized electrodes 1 cm. in diameter and 8 cm. apart. Cell C was 8 cm. longer than cell B, and the electrodes were 8 cm. further apart; the cells were otherwise identical. Cell D was designed according to Jones and Bollinger.lo The electrodes were unplatinized and their area and the distance between them were approximately the same as for cell A. An oil thermostat controlled the temperature to 0.01".
tertiary butyl alcohol isolated, dried, and identified by melting point and the melting point of a mixture with a n authentic sample. Tertiary Butyl Betainium Chloride.-Twenty grams of tertiary butyl chloroacetate and 62 cc. of a 2.5 molar solution of trimethylamine in dry benzene were mixed with cooling. Although crystals formed in a few minutes, it proved advisable to allow the mixture to stand for twenty-four hours; yield, 18 g. The crude material can TABLEI be crystallized from acetone or dioxane to which a small RATEOF CHANGE OF CONDUCTIVITY OF SODIUM HYDROXIDE amount of water has been added. The product is a OF GLASS SOLUTIONS DUE TO SOLUTION white crystalline powder, which, when dry, melts with Alkali concn., Change in % per hr., a t decomposition around 220", the exact temperature deCell AT 30' 350 pending somewhat on the rate of heating. A 0.002 0.022 Anal. Calcd. for &H2oO2NC1: C1, 16.91; N, 6.68. B ,005 .064 Found: C1, 16.89; N, 6.79. 0.10 B .008 The material used in the rate determinations had been B .030 .022 recrystallized several times, and was of analytical purity. C ,015 .030 .050 Tertiary Butyl Dimethylg1ycinate.-Twenty grams of C .050 ,012 tertiary butyl chloroacetate and 43 cc. of a 4 molar soluMaterials.-Tertiary butyl acetate was prepared accord- tion of dimethylamine in dry benzene were mixed, with ing to Skrabal and Hugetz," oxamide according to LiebigiZ cooling, and allowed to stand a t room temperature for and oxamic acid according to Oelkers.l a Dimethylglycine twenty-four hours. The resultant material was extracted was synthesized by the method of Michaelis and Schubert14 with 1:1 hydrochloric acid, and the extract washed once and the chloride of betaine amide by the method of Ren- with ether. The aqueous layer was then saturated with shaw and Hotchkiss.16 To determine its purity, each was potassium carbonate and extracted with ether. The ether analyzed before use. extract was dried with potassium carbonate, the ether reTertiary Butyl Ch1oroacetate.-While the tertiary butyl moved by distillation, and the residue fractionated a t rechloroacetate used in the velocity determinations was pre- duced pressure. The yield of colorless oil, having a slight pared by direct esterification, a more convenient method ammoniacal odor, was 12 g. The hydrochloride was preof preparation follows: 60 cc. of chloroacetyl chloride was pared by adding a slight excess of hydrogen chloride in placed in a three-necked flask equipped with thermometer, methanol, and evaporating to dryness under reduced stirrer, dropping funnel and an outlet protected by a cal- pressure. The residue can be recrystallized from a 10% ciuni chloride tube. A mixture of 100 cc. of dimethyl- solution of tertiary butyl alcohol in benzene. The white aniline and 75 cc. of tertiary butyl alcohol (m. p. 23') was twice recrystallized material which was used in the vcadded dropwise during the course of an hour, maintaining locity determinations melted about 150°, the exact temthe temperature a t 0" by means of an ice-bath. At the perature depending somewhat on the rate of heating. end of this time, the ice-bath was removed and the mixture Anal. Calcd. for CaHls02NCl: C1, 18.13; N, 7.16 allowed to stand overnight. Water was then added, and Found: C1, 18.08; N, 7.12. the heavy ester layer separated, washed with water, dilute Conductivity water and carbonate free alkali were used hydrochloric acid, sodium bicarbonate solution, and water throughout. Chemicals other than those recorded here again. The oil was dried over calcium chloride and were of reagent grade. vacuum distilled: yield, 72 g., or 60%, of the ester. The Method.-The usual procedure in the deterproduct is a colorless liquid, slightly heavier than water, boiling a t atmospheric pressure a t 155' with some de- mination of reaction velocity by conductivity ~ composition and a t 60.2" a t 15 mm., n Z o1.4230. is to determine the initial and final conductivities Anal. Calcd. for CcHliO2Cl: C, 47.8; H, 7.37; C1, of the solution and to calculate the rate on the 23.56; sapon. equiv., 150.5. Found: C, 47.6; H, 7.33; assumption that the conductivity of the solution C1, 23.63; sapon. equiv., 150.6. is a linear function of the extent of reaction. In The material actually used in the velocity determinations the determinations here described, since it was had been carefully fractionated a t reduced pressure, and inconvenient in most cases to measure the initial was of analytical purity. An acidified portion of completely saponified ester gave and final conductivity, each velocity determinano precipitate with silver nitrate, showing that there is no tion was standardized by measuring the conappreciable hydrolysis of the chlorine during saponificaductivity of solutions corresponding in ionic tion. Another portion of the ester was hydrolyzed, the (10) Jones and Bollinger, THISJOURNAL, 53, 411 (1931). (11) Skrabal and Hugetz, Monafsh., 4'7, 17 (1926). (12) Liebig, A?z;z., 9, 1 (1834); see "Beilstein," IV ed., Vol.
11, p
,546.
(13) Oelkers, B e y . , 22, 1566 (1889). (14) Michaelis and Schubert, J. Biol. Chem., 115, 221 (1938). 48, 2698 (1926). (1.5) Renshaw and Hotchkiss, THIS JOURNAL,
composition to that present a t several different times during the reaction.16 Since the concentrations of the various ions in the standardizingsolution could not conveniently be made exact (16) A similar standardization procedure has been described by Crocker and Lowe, J . Chcm. Soc., 91, 952 (1907).
F. H. WESTHEIMER AND MARTIN W. SHOOKHOFF
272
duplicates of those at any point during the reaction, a small correction had to be added to the conductivity of each standardizing solution. In order to estimate this correction, i t was only necessary to know the ratios of the conductivities of all the ions present to that of hydroxide ion.
